This Project will compare healthy aging and senile dementia of the Alzheimer type in a sample aged 80 and older with respect to longitudinal clinical and psychometric performance and neuropathologic data postmortem. Guidelines for the separation of Alzheimer and nonAlzheimer brains on the basis of quantitative neuropathologic markers in the very old will be developed. The extent of overlap in the brains of demented and nondemented very old persons will be examined. In the process we will extend our knowledge about clinical and psychometric performance in a nondemented sample of persons aged 80 and older. We will have studied approximately 400 very old subjects during life and expect to have at least 125 autopsies from this sample at the completion of this proposed Project. About one-third of the brains are expected to come from the nondemented and mildly/very mildly demented (SDAT) and two-thirds from the more severely demented very old subjects.